$100M isn’t mega anymore, Waymo in China, and giving AI curiosity 🔍
August 26, 2018
The AI and Robotics Newsletter from Comet Labs. Subscribe here. Past issues here.
👋 Hi!
We have some exciting news coming soon with a couple of new investments (!) and our next Assessing China program focus. If you missed out on our last one, be sure to keep an eye on our website for the details and how you can apply!
In other news, while OpenAI Five was ultimately defeated in its two games at The International, it was very fun and interesting to follow along with the updates and see their progress. It did get us thinking about the broader implications of the technology in OpenAI Five and how it’ll be used beyond DotA. Let us know your thoughts below in this week’s lunch table poll.
Have a happy Sunday and thanks for reading!
Follow Comet Labs for more stories and thoughts about the impact of AI and robotics.
💬 THIS WEEK AT OUR LUNCH TABLE:
Given the OpenAI Five is good enough now to compete at the highest level of professional DotA, what does their progress mean to you?
Amazing! The system can learn in complex environments! |
Great, but let’s test in other places, like driving. |
Why esports? Focus on real-world problems! |
I don’t follow esports or DotA.Last issue, we asked you about which emotions you think we should train algorithms to detect best in interactions with humans. Many of you agreed that Anger was the most important emotion. We’re not surprised, given the explosion of AI as a customer service assistant!
📚 FEATURED ARTICLE:
$100 Million Was Once Big Money for a Start-Up. Now, It’s Common. — New York Times
With mega funds like SoftBank’s Vision Fund encouraging increasingly larger check sizes, startups now have access to more capital than ever. But while more money gives founders speed advantages, raising too much is still a risk.
📹 WATCH THIS:
A team from Nvidia and MIT have created a video synthesis system that can replace video footage with parts of another video at a surprising level of accuracy above what was previously possible.Read their paper here.
📖 WORTH READING:
China Mines Silicon Valley for Tech Talent — Wall Street Journal
While the US is taking a defensive stance, China’s business leaders and entrepreneurs are eager to attract skilled employees, particularly engineers.
What algorithmic art can teach us about artificial intelligence — The Verge
Some kinds of art aren’t supposed to look or feel the same even between humans. Add an AI into the mix, and what does it see?
AI is the future — but where are the women? — WIRED
Diversity and inclusion have never been great in the computer science field, but it’s even worse specifically in the machine learning research community. Several organizations, like Women in AI and Black in AI, are working hard to change that.
What’s the Purpose of Companies in the Age of AI? — Harvard Business Review
Following the assumption that AI capabilities will continue to increase exponentially and augment/replace tasks performed by humans, it may be possible that structured companies may no longer be necessary.
Give AI curiosity, and it will watch TV forever — Quartz
This is beyond TV. Understanding curiosity and effectively recreating it within AI is another step toward applying algorithms to help discover things that we can’t on our own. This includes generative design, new materials, biological simulation, and much more.
Finding the Goldilocks Zone for applied AI — TechCrunch
Ivy Nguyen from Zetta Ventures wrote a nice piece about being careful in designing an AI strategy for a business challenge. There are right and wrong times to adopt it or to launch an AI startup.
How 5G connectivity and new technology could pave the way for self-driving cars — MIT Technology Review
For consumer AI technology to reach its full potential, especially with self-driving cars, we’ll need the speed and efficiency of 5G technology. Luckily, it’s closer than it may appear.
Waymo Sets Up Shanghai Subsidiary As Google Makes Inroads In China — China Money Network
In addition to the recent controversial news that Google has started work on a censored search product specifically for China, Waymo has also made steps toward setting up in the Chinese market, further positioning Google as a competitor to the Chinese technology giants.
📝 READ THIS PAPER:
Opening the ‘black box,’ Google DeepMind AI system diagnoses eye diseases and shows its work — De Fauw, et. al
Researchers from DeepMind recently announced that it had created an algorithm that is able to detect specific eye diseases at a level of accuracy equal to eye specialists. The system was able to make the correct recommendation in 94% of cases it faced. In their paper, the DeepMind team points out the rapidly increasing volume of OCT scans is not being met fast enough by trained human experts who can interpret them, presenting a problem that an AI solution is specifically capable of assisting with. Importantly, the DeepMind team also created a framework that matches the human decision-making process that separates individual judgments the algorithm makes, breaking down the “black box” problem and allowing the physician to interpret all parts of the algorithm’s recommendation. Read the full paper on Nature here.
👏 PORTFOLIO NEWS:
Ed Husic’s AI centre of excellence to focus on ethical, humanist AI — Akin
The new AI center in Australia will examine the ethical implications of AI, teaming up with AGI research company Akin to make it happen.
Dezeen Awards 2018 design shortlist showcases the hottest new products and concepts — Lightform
The 2018 Deezen Awards have shortlisted Lightform in their Digital Design category! Awards will be announced on November 27th, so stay tuned.
🛠 STUFF YOU CAN USE:
Video-to-Video Synthesis — Nvidia, MIT
Check out the full paper and GitHub repo from the team from Nvidia and MIT who created the video synthesis engine shown in the video above.
Writing with the machine — Robin Sloan
This clever tool created by author/programmer Robin Sloan is an autocomplete engine for writers. Trained on sci-fi text, it will simply analyze new text inputs by a human and generate new text to try and create compelling stories. Sloan talks more about his invention and the intersection of AI and creativity with ArsTechinca here.
100 Days of ML Code Infographics — Avik Jain
If you’ve never coded in Python, tried your hand at ML or data science, or you simply want a refresher on the fundamentals, Avik Jain created a GitHub repository of helpful infographics (inspired by 100 Days of Code) that communicate a remarkable amount of helpful knowledge very quickly. Check them out!
📣 HIRING:
Mobile Engineer — doc.ai
The team at doc.ai is on a mission to revolutionize the use of healthcare data with blockchain and AI, and they’re looking for a mobile engineer to join the team.
Several Positions Available — Creator
From engineering to design, on both hardware and software teams, Creator is looking for talented folks for many different roles.
Several Available Positions — Grabango
Grabango, our portfolio company providing technology for checkout-free shopping, is looking for new members to join its stellar team.
And that’s it for this week! Find us on social media to stay updated with the latest news in AI and robotics.
